[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: time for a new cvs-1.12.x release?

From: Paul Edwards
Subject: Re: time for a new cvs-1.12.x release?
Date: Fri, 26 Sep 2003 22:25:46 GMT

"Derek Robert Price" <address@hidden> wrote in message news:address@hidden
> I'm for a release of stable, but there is still a bug that looks like a
> race condition only on Mac OS X that I would like to see fixed on
> feature.

Or perhaps, YOU've only been able to produce the problem
on Mac OS X, others may be unfortunate enough to be able
to produce it on ...

> There is a consistent failure of the checkout_repository-1
> test, but when run via the debugger, the test fails to occur.  Mark

Sounds more like a wild pointer then.  Just waiting for the
right combination of compiler, library, OS etc to rear its
ugly head.  Fortunately its reproducible on one of those

> mentioned that it seemed to be dependant on which tests were run before
> it as well.
> As for hurrying to get patches into a release, that way lies madness.
> Or at least hurried & broken code which leads to bugs in releases.  :)
> I'll think about a stable release sometime next week if nobody has a
> good reason why I shouldn't release it.

Great - thanks!

> I'd really hate to release feature with a known OS X bug, but I don't
> have time to figure it out just now and I may consider releasing it with
> a warning for users of Darwin & OS.  It'd be good to get out Brian
> Murphy's PAM stuff and Mark's BSD-CVS compatibility stuff.  Still, I
> don't want feature to get that kind of dirty reputation for bugs and I'd
> like to hear some discussion.

It will get exactly that dirty reputation if you release it with a
potential wild pointer, known beforehand and readily
reproducible.  Remember your quote above about "that way
lies madness"?  :-)

BTW, when it is found, I'd be interested to know if the bug
was introduced by new code or whether it was an existing

BFN.  Paul.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]